World Health Organization

WHO1BS108.2092 ENGLISH ONLY

EXPERT COM31ITTEE ON BIOLOGICAL STANDAmIZATION Geneva, 13 to 17 October 2008

Extended use for the International Standard for Alpha-l-antitrppsin (051162):

- To assign potencies to recombinant products - To assign a total protein and antigenic concentration

Craig ~helwell', Ewa ~ a r s z a l " Peter ~ i g s b ~ h n d Coiin ~ o n ~ s t a f f '

Biotherapeutics Group, Huemostasis Section' and Biostatistics ~ r o u ~ ' Rrational Institute for Biological Standards and Control,

Blanche I~rze , South iwi~nms, I'otters Bar, Herts, Ely6 3QG, C'K

ljivision of Hen~atolo~'?, FDMCBER, 29 Lincoln Drive, Betlzesda, itIIj20892, USA

O World Health Organization 2008

All rights reserved. Publications of the World Health Organization can be obtained from WHO Press, World Health Organization, 20 Avenue Appia, 121 1 Geneva 27. Witzerland (tel.: +41 22 791 3264: fax: +41 22 791 4857: e-mail: bookorders@who,int). Requests for permission to reproduce or translate WHO publications - whether for sale or for noncommercial distribut~on - should be addressed to WHO Press, at the above address ifax: +41 22 791 4806; e- mail: permissions@who,int).

The designations employed and the presentation of the material in this publication do not imply the expression of any oplnion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines on maps represent approximate border lines for which there may not yet be full agreement.

The mention of speciiic companies or of certain manufacturers' products does not imply that they are endorsed or recommended by the World Health Organization In preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters.

All reasonable precaut~ons have been taken by the World Health Organ~zatlon to verify the ~nformation contained in thrs publ~cation However the published material is being distributed without warranty of any kind either expressed or lm~l led The responsibility for the interpretation and use of the material lies w~th the reader In no event shall the World Health Organization be liable for damages arisrng from its use

The named authors [or editors as appropriate] alone are responsible for the views expressed in this publication.

WH01BSfO8.2092 Page 2

Summary The 1'' International Standard (IS) for Alpha-l -Antitr\;psin (AAT) was established by the World Health Organisation (WHO) Expert Committee on Biological Standardisation (ECBS) at their annual meeting in Geneva 21-27 October. 2006. The preferred candidate 051162 was established as IS with an assigned potency of 12.4 n ~ g 1243 nmoles) active AAT. Use of the standard was restricted by the ECBS to plasma-derived products.

In this report we present the findings from a follow up coliaborative study that was organised to investigate the use of the IS for assigning potencies to recombinant AA7 preparations. The study involved 5 participants from 3 different countries. Laboratories were asked to measure the potency of 5 .4AT preparations against elastase using the IS (0511 62). The samples included the plasma preparation 0511 50 (candidate A in the first potency study') as an internal control alongside recombinant and plasma products anonyn~ously labelled E. F. G and H. Data analysis indicated no differences between the results obtained for plasma-derived products and the recombinant products from yeast, transgenic goat and transgenic sheep. It is therefore proposed that the use of the I " IS for AAT be extended fhr potency determination of recombinant products with a potency value of 233 nmoles, in addition to the plasma-derived products already approved.

The potential use of the AAT IS (0511 62) as a standard for total protein and antigen determination was investigated by a subset of the participants of the first potency study'. The results were presented in the previous study report. however due to the overall variation in the results between the different methods and internal standards used. it was proposed in agreement with the study participants that further work was required to assign a value to the standard with confidence. In this report we present the results of independent amino acid analyses that support assigning a total protein value to the AAT IS (0511 62) of 12.4 mg, consistent with the potency value. The results of the total protein and antigenic assays from the previous study were re- calculated based on this value and the variability of the results was significantly reduced. Based on these results we propose that the AAT IS (051162) is assigned a total protein and antigenic value of 12.3 mg.

Introduction Alpha- l -Antitrypsin (AAT), also known as Alpha- l -Proteinase Inhibitor (A I -PI). is a member of the serpin (serine protease inhibitor) family and its physiological target is elastase. Alpha- l - Antitrypsin Deficiency (Alpha-1) is a genetic disorder that can cause liver and lung disease in adults and children. Plasma-derived AAT therapeutic products are used to treat Alpha-I and the requirement for a reference preparation to standardise the measurement of potencies of therapeutic products of AAT was recognised by regulators and industry. Consequently in 2005106 an international collaborative study was organised1 to establish such a reference preparation. Four candidate AAT preparations (3 plasma-deritfed. one yeast recombinant) were used in the study labelled A. B, C and D and the details of these are provided in the materials section below. Potencies tvere assigned to each of these preparations based on the results of the study in which each AAT candidate was titrated against active-site ritrated trypsin. Potencies were assigned in terms of nmoles and milligrams of active inhibitor. in line with the labelling practice of licensed AAT therapeutic products. and candidate C 1051162) was chosen as the standard in agreement with the participants of the study.

The result5 of the study \bere pretented to the WHO Expert Committee on B~ologlcal Standard15atlon (ECBS) and the U'HO ~'"ntemattonal Standard (IS) tor AAT (051 162 wai, establlrhed at their meeting In Geneta, October 2006 The IS ua5 approted for asslgnrng

Page 3

potencies to plasma-derived AAT therapeutic products and the report of this study was prrblished as a WHO ECBS technical report'. This doc~~ment is referred to throughout this report and is reproduced in Appendix D.

With a number of different recombinant therapeutic products in de~elopment by several different manufacturers it is important that these products are labelled with potencies in line with the licensed plasma-derived products. A second collaborative study was therefore organised to deterntine if the IS (051162) is suitable for this puvose. In this study, presented here in part 1 of this report. recombinant products were tested in assays together with plasma-derived products anonymously labelled E, F, G and H (see the materials section below).

A further potential use for the IS (051162) is as a total protein and antigenic standard to determine the specific activity of AAT preparations. As such a subset of the participants of the previous potency studyL carried out total protein (Bradford) and antigenic assays on each of the four candidate preparations A. B. C and D. The results of this were presented in the potency study report'. however due to the variation in the results between the different labs and the different standards used. it was felt that values could not be assigned to the preparations with sufficient confidence and it was therefore proposed in agreement with the study participants that this should be investigated further. In part 2 of this report we present additional work carried out to support assigning a total protein and antigen value to the IS (051162).

Materials A summary of the AAT preparations used in this work is provided in the table below:

AAT sample I NIBSG code I Source International Standard 1 0511 62 I Human olasma

1 D 1 0511 72 I Recombinant I l

L

E I PM-07-01 3 I Recombinant F I

1 051 1 72 I Recombinant G 0511 50 l

Human Plasma H

1

l 991696 1 Recombinant

Detailed information of the source and fills of samples '4. B. C and D can be found in the l previous potency study report , Two additional recombinant preparations were used in this

work: a transgenic AAT from PPL Therapeutics. Edinburgh. UK. expressed in and purified from sheep milk. This material was from a trial fill carried out at NIBSC in 1999 and the product is no longer in development. The second material was also transgenic, but expressed in and purified from goat milk. This material was provided by GTC Biotherapeutics, Massachusetts. USA and filled at NIBSC in 2007. In addition. sample '4 10511 50) was re-coded as sample G which was assayed alongside the recombinant products anonymously.

Human plasma A

In the previous potency study reporththe predicted stability of each of the candidate samples A (051150) B (051152) C (0511 62 ) and D (0511 72) was calculated based on the results of accelerated degradation trials. In ilppendix A of this report we present an update of predicted stabilities based on new data.

B 0511 52 l Human plasma C 1 0511 62 I Human plasma

0511 50

WHOfBSIO8.2092 Page 4

PART 1 Assigning potencies to recombinant AAT preparations A small collaborative study was organised to investigate whether the IS for plasma-derived A,AT (0511 63) is appropriate for assigning potencies to recombinant ArZT preparations. It is important that reconlbinant AAT products which are under de.ilelopment are assigned potencies that are consistent ~ l i rh the plasma-derived prodcrcts currently Iicensed and in development. The aim of this stud!; is to determine if the I " IS for plasma-derived iZAT (051162) is suitable for this purpose.

Study design Five panicipants were invited to take part in the stud!;. 3 from regulatory Laboratories and 2 from manufacturers of recombinant AAT products (US (3). UK and Germany}. The list of participants is given in Appendix B. Each laboratory was assigned a code number and the order of listing does not correspond to the identifying number assigned to each laboratory.

Each participant was provided with 4 ampoules of 6 different AAT preparations, including the IS (051152j and candidate A (0511 50) from the first potency study1 as an internal control. The remaining 4 preparations were anonymously labelled E. F, G and H. and included 3 recombinant products and an additional ampoule of the plasma-derived A (0511 50). The recombinant proteins were from differerlt sources tyeast, transgenic sheep and transgenic goat produced in milk) to represent a range of different products. This information was not provided to the participants. although an approximate potency of each unknown and a suggested pre-dilution series for each was given.

Each participating group was requested to perform a minimum of 3 assays titrating each AAT preparation against elastase using their usual in-house method. In the first potency study' active- site titrated trypsin was used as the inhibitory target which allowed the direct measurement of AAT potency to assign a value to the IS. AAT potency measurements can now be made relative to the IS by titrating against elastase, the physiological target for AAT. This is the recommended approach for potency determination by manufacturers and forms the basis of the European Pharmacopoeia method due for publication July 2008. The principle of the assay is to titrate a range of dilutions of each AAT (minimum 3) against a fixed concentration of elastase. Residual elastase activity is measured using a suitable chromogenic substrate for elastase. monitoring the change in absorbance over time as either a kinetic rate or in an endpoint assay. A method was provided by NIBSC to groups upon request.

Results and discussion Assay results were returned by all laboratories to NIBSC for analysis. Lab I returned results for 4 kinetic assays, lab 2: 3 endpoint assays, Lab 3: 3 endpoint assays, lab 3: 3 kinetic assays and lab 5: 3 kinetic assays. Potencies and confidence limits for the internal control A i051150) and each of the test preparations E. F, G and W were calculated relative to the IS (051152) using the slope ratio modet ~ l i t h a completely randomised design. Calculations were made using the Combistats programme, software developed by the European Directorate for the Quality of itledicines and Healthcare (EDQM) that can perform calculations according to the European ~harmaco~oeia'. This is with the exception of laboratory 4, whose data included a single dose for the standard and test preparations for each assay. In this case potencies were assigned to the test preparations as ratios relative to the result of the IS. The results for samples h, E. F, C and H are given in Tables l . l - I .5 and Figures 1. l - 1.5 respcti?fely. For each sample the laboratoy arithmetic mean potency and standard deviation is shown together with the combined mean, inter-laboratory standard derriation and coefficient of variation (CV 5%) . The results are summarised in Table 1.6 including the calculated 95 % confidence internals.

WHOIBS108.2092 Page 5

The aim of the potency resting of the different AAT preparations was to det-emine if the IS is suitable for use in assigning potencies to recombinant A.Yf products. Candidate A (0511 50) was used as an internal control and a mean potency of 203.01 nmoles was calculated. This value is within 2 72 of the potency assigned using active-site titrated trypsin in the first potency study" 1205.0hmoles). 051150 was re-IabelIed as sample G anonymously in the study and the results were consistent t;rfith sample A with a potency of 211) .23 nmoles. Sample F in the study was 051172. a recombinant preparation that %as included in the first potency study' as Candidate D. In that study 051172 was assigned 3 potency of 197.25 nmoles against trypsin. The result of 183.7 1 from the present study is within 7 % of this value and is not significantly different (analysed by an unpaired t-test, P-Value = 0.183)' The two remaining recombinant preparations, E and H. were assigned potencies with cornparable CVs to the other materials, with all CVs being below 10 %. No s~stematic deviations were observed in the analysis of any of the materials in the assay and no differences were observed between the recombinant and the plasma-derived makrials.

Assay va l id i~ Multiple dose assays analysed by Combistats were considered 'statistically valid' if the outcome of the analysis of variance met the following criteria. as described in the European ~hamaco~oe ia ' :

I . The variation due to blanks (elastase with no AAT) was not significant, i.e. the calculated probability was not less than 0.05. This indicates that the responses of the blanks do not significantly differ from the common intercept and the linear relationship is valid down to zero dose.

2. The variation due to intersection was not significant i.e. the calculated probability was not less than 0.05.

3. In assays including at least 3 doses per preparation, the variation due to non-linearity was not significant i.e. the calculated probability was not less than 0.05.

In cases where one or more of the above criteria were not met. i.e the calculated probability of variation due to blanks, intersection or non-linearity was less than 0.05, the assay results were scrutinised further. The typically high precision of this type of assay sometimes results in apparently significant variation in the validity criteria. In such cases a visual check of the dose response curve for each preparation in an assay was carried out to cheek for any systematic deviation. Linearity was then confirmed by excluding the highest dose from the analysis checking for a significant increase in the caiculated potency. For the assay to be considered valid at least 3 doses must be included for both the standard and the test preparation and any deviation from the validity criteria deemed incidental rather than systematic.

The -caIldrt) cntena for the mgle dose rewlts from lab 3 \%ere for an ~nhlbttlon rewlt ot betueen 30 (?c and 70 % of the elastase on15 (no AAT) blanks.

No rerults were excluded h m the analysis baied on the acceptance criteria outlined abote

Conclusion Rased on the re\ult\ of t h ~ s rrudq ue conclude that the IS 1s surtable for a5wgnlng potencies to recomb~nant AAT products u tth the stated potency of 243 nmoles.

WHOBS108.2092 Page 6

The milligram potency must be converted from the derived potency in moles for individual recombinant products from different sources, based on actual molecular t-tieights for each protein. This advice is incltlded in the IFC (Appendix C!.

Proposal IFU be modified to remove rei*tncttons on the 1" IS for AAT 1051162) 50 ~t can now be used as an IS for recctmb~nanrltranclgen~c product\.

WHO1EfSfOS.2O92 Page 7

Table 1.1. hlean potencies of sample '4 (nmoles AAT per ampoule). Potenciei are caiculated from the rnean reiult4 from all the ai ials returned.

Figure 1.1. hiean potencies of sample A (nmoles AAT per ampoule).

400 7 AAT potency of sample A (Q511 50)

Mean

- - - - - - - - Confidence limits (95 %)

Lab number

WHO1BSfO8.2092 Page 8

Table 1.2. %lean potencies of sample E (nmoles AAT per ampoule). Potencie., are calculated from the mean reiult\ tram all the as\ay\ returned.

Figure 1.2. Rlean potencies of sample E (nmoles AAT per ampoule).

AAT potency of sample E

Mean

- - - - - - - - Confidence limits (95 OA)

Lab number

WHOIBSfO8.2092 Page 9

Table 1.3. Mean potencies of sample F (nmoles AAT per ampoule). Potencies are calculated from the mean results from all the assays returned.

Figure 1.3. Mean potencies of sample F (nmoles AAT per ampoule).

AAT potency of sample F

Lab number

WHOIBSIO8.2O92 Page 10

Table 1.4. Rlean potencies of sample G (nmoles AAT per ampoule). Potenc~es are calculated from the mean re\ult\ from all the as\ay\ returned.

Figure 1.4. Mean potencies of sample G (nmoles AAT per ampoule).

AAT potency of sample G

Mean

- - - - - - - - Confidence limits (95 %)

Lab nurnber

\VH01BS/OS.2092 Page 1 l

Table 1.5. Mean potencies of sample W: (nmoles AAT per ampoule). Potencies are calculated from the mean rewlti, from all the a\i,ay\ returned.

Figure 1.5. Mean potencies of sample H (nmoles .4AT per ampoule).

AAT potency of sample H

- 'loo at B

3 g 80 E a I

60 B

0 E 5 40

Mean

- - - - - - - - Confidence limits (95 %)

Lab number

\vH01BS/O8.2092 Page 12

Table 1.6. Summar) of the mean potencie5 for each of the candidate\.

1 AAT sample iNIBSC code) I (nmo

PART 2 IS (051162) as a total protein and antigenic standard There has been a desire expressed by manufacturers of therapeutic AAT products for a total protein and antigenic standard for AAT. It was proposed that one of the candidates in the first potency study' could sen7e as the standard and this was investigated by a subset of the study participants. Bradford assays and antigenic assays were performed by these laboratories using their own methods with in-house standards and a highly purified AAT standard from Athens Research Technology (ART), which was provided to all groups in the study. The results for this were presented in the potency standardisation study report' separately for the ART standard and the laboratories' in-house standards.

The combined results for total protein by Bradford assay (table 2.1 ) and for antigenic determination (table 2.2) for the three plasma-derived candidates (A (0511 50), B (0511 52) and C (IS) (051162)) are reproduced from the first study' below.

It was agreed by the potency study participants at the time" due to the variability of the data, that further investigation was required to assign values to the preparations with confidence. In this section we now present additional data to support assigning total protein and antigen values to the IS (0511 62 f .

WHOIBSIO8.2092 Page 13

TabIe 2.1. Total protein concentration determined by Bradford assay using difkrent standards. results reproduced from the potency study1.

Candidate

3 1 AAT 1 11.33 1 4 / ART 1 11.93 1

l

9 I AAT I 14.56 1 11.09 1 "ART h~ghly purified AAT standard from Athen\ Retedrch Tecnnologiet

Table 2.2. Antigenic determination of AAT potency study candidates using different standards. results reproduced from the potency study'.

".ART: highlq purified AAT standard from Athen. Ke>earch Technoiogies

WHOIBS108.2092 Page 14

Results and discussion

Atnino acid atzalysis of IS (05/162). Amino acid analysis is an absolute method of detemining protein concentration. The process involves protein hydrolysation Followed by ion exchange separation of the amino acids foIlovr.ed by post column photometric detection. This approach is available to us since the IS 051162 was filled as purified AAT without additional protein excipients. Amino acid analysis was carried out by two independent laboratories: Afta Bioscience, Birmingham, I;K and the Protein?; and Nucleic Acid Chemistry (~PNAC) Facility, Cambridge, UK. For each a single ampoule of OS/ l62 was reconstituted and split into two fractions. Analyses were carried out on each fraction and the results were reported in terms of mg and nmotes of amino acids per ampoule.

The table below gives the replicate results received from Alta Bioscience and PNAC Facility in terms of m ~ m p o u l e . The mean result is shown and converted to nmoleslampoule using a molecular weight of 44.5 kDa, based on the amino acid composition.

Table 2.3. Total protein results for IS (051162) by amino acid analysis.

I / Sample l 1 Sample 2 Rilean result 1 Mean result 1

The accuracy of the method is reported to be. on average, +S 5%. While most of the amino acids are stable, the hydrolysis conditions do result in some losses. 10 5% of serine and 5 C / f of threonine is lost and this is accounted for in the calculation. All tryptophan is destroyed. which represents 2.32 9 of AAT by weight, and this was also corrected for in the calculation. Losses of cystine are variable. however this accounts for only 0.46 9 by weight of AAT and so was ignored.

img/ampoulc) (mg/ampoule)

Alta Bioscience I 10.08 9.63 9.94

The total protein content of AAT IS (051 162) can therefore be taken as 223.37 nmoleslampoule 5.5 $4 (21 2.20 - 233.54). or 11.39 mglampoule ( 1 1.46 - 1 1.96) based on a molecular weight of 5 1 kDa. corresponding to the glycosyiated protein as determined by LClMS (see previous report!). This is compared to the stated potency of 243 nmoleslampoule (231 - 255), or 12.4 mglampoule (1 1.8 - 13.0) where the range corresponds to the 95 C?c confidence interval in the previous study1.

(nrnolc\/ampoule)

The calculated total protein concentration of 1 1.39 mg/ampoule is approx 8 Cic lower than the potency value ( 1 2.4 mgampoule active AAT) determined in the first potency study' and the total protein value determined by Bradford assay is approx 5 5% higher (13.01 mglampoule). These values are however all within the 95 Cir confidence intervals stated for each:

l 223.37

PNAC Facilitv 10.02 10.03

WHOlf)S/08.2092 Page 15

Table 2.4. Potency, antigen and total protein results for IS (051162)

1 121210 113 531 i 1146 -11961 i

Total proteln (Bradford atsay - table 3.1 ) l I

l 255.10 I ---7ET+ I

j nmoles j n~ g ; / , i95 C; confidence interhaii I (95 ''; confidence interval I ,

In the previous potency study report' one parricipant commented in their response to the proposals that because of the high purity of candidate C (051 162) (IS), this could be assigned 100 52 antigenic concentration. The additional data presented here supports this and since the total protein concentration can't be lower than the potency. we propose that a total protein and AAT antigen value of 243 nmoles1ampoule ( 12.3 mg for plasma-derived products) be assigned to IS (051162). Assigning a total protein equal to the potency is in line with the intermediate value of the 3 shown in table 2.4, and the 95 % confidence intervals are overlapping.

, Potency (from previous study1) i I Total protein (~arnino acid analysis)

Re-calculating the total protein and antigenic results using 051162 as the standard If the IS (0511 62) is assigned a value for total protein of 12.4 mg as described above. the results from each of the laboratories' assays above (from the previous study1) can be re-calculated using this standard value. Table 2.3 below shows the re-calculated Bradford assay results from table 2.1, and table 2.4 shows the re-calculated results of the antigenic assays from table 2.2. Table 2.5 shows a summary of the total protein, antigen and potency results for A (0511 50) and B (051152) with the 95 % confidence interval for each. All of the original data was from the first study'.

Fewer results are shown in tables 2.5 and 2.6 than in 2.1 and 2.2 above. This is because in some cases the results from a single assaf were analysed by more than one standard. For example. laboratory 3 analysed a single data set with AAT, albumin and human serum standards and these results were presented separately. When the result was re-calculated with IS (051162) as the standard only one result is obtained. The data in tables 2.3 and 2.4 therefore represents the re- calculated values of each individual assay performed.

233.12 12.30 1211.21 - 255,03) I i l l . " - l3.Oli

The antigenic results for laboratory 13 (table 2.7) were greater than two standard deviations from the overall mean values and were therefore excluded from the calculation as statistical outliers. The result for IS (051162) seemed particularly low in this assay. therefore the results tbr A i051150) and B (0511 52) appeared to be overestimated.

223.37

The \ anab~l~ t ] of the data for both the total protetn and antigenic retults 1s greatly reduced when the IS 15 uted as the standard, ~ndicated bq the loner CV saluet In table 2.3 (4.17 % for A (0511 50) conlpared to 13.61 4 and 6.29 Q for B (0511 52) compared to 13.55 Cic) and table 2 3 ( 1.81 Cic compared to 1 1.3 for A (05f150) and 3 7 c> compared to 8.91 4; tor B (051152)).

l 1.39 I

FVHOIBS108.2092 Page 16

Table 2.5. Total protein deternlination by Bradford assay using the AAT IS 1051162). as the standard ( I 2.3 mg).

Table 2.6. Antigenic detemination of the potency study candidate\ using the AAT IS (0511 621, as the 5tandard (1 2.3 mg).

*The results from invalid assays. indicated by brackets ( ). were not included in the calculation.

Table 2.7. A summaT of the total protein and antigen results using on the IS (051162j as standard. compared to the potency results for samples A (0511 50) and B (051152). The results are given as mglampoule.

Conclusions The total protein results described above. incIuding the Bradford assays and the amino acid analysis. are consistent with the potency value already assigned to 0511 62 suggesting a specific activity close to 100 % with the expected errors of the methods used. We therefore propose a total protein value of 233 nmoleslampoule (1 2.4 mg for plasma-derived products! is assigned to the 1" MiWO International Standard for AAT (051162) in line with the potency value.

WH01BS/O8*2O92 Page 17

Ba\ed on the reduced vanabll~ty of the rewlt5 with the total protein and antlgenlc assays \%hen re-calculated ~ i t h the IS (0511 62) a\ standard we propose that that the IS 10511 62) potencj valrre of 12 3 mg can he used to awgn total proteln and antigentc value\ to AAT preparation\.

For antigenic assays we propose including a disclaimer in the instmctions for use that individual laboratories should investigate the suitability of the antibody used for the material to be tested. This is pafiicularly imporlant when recombinant products are being tested where differences in glycosylation patterns may affect the results.

Proposals The 1" IS for AAT (051162) be a5signed a total protein and antigenic value of233 nmoledampoule i 2 2.3 mglampoule).

References I . International collaborative study to establish the WHO ~'"nternational Standard for

Alpha-l-Antitrypsin. (051162). C. Thelwell, P. Rigsby and C. Longstaff. WHO Expert committee on biological standardization technical report. WHO/BS/06.20.F4.

2. Statistical analysis of results of biological assays and tests. 01/2008:50300 European Pharmacopoeia 6.0, EDQM; Section 5.3. p571

Participantsf response Participants involbed in all parts of the study were asked if they agreed or disagreed with the folio% ing proposals:

I . We propose rhut the l"' Inter~zational Sturzdard.for Alpha-I-Aiztitvpsin (05/162) is snituhle f i r assigning poteacy i~alues to recornbinurzt .&IT products with the stated potency of 243 nrnoles active AAT per arrzporlle.

2. U7e propose a total protein value qf243 nmole.r/c.lmpoule t 12.4 nzg,fir plu.sr?za-derived prc-rdztcts) is ussigvzed to the l" WHO International Stanllurd,for AAT (05/1621 in li~ze 1t.ith tlzr potenq valrte.

3. We prcFore arz antigerz ~~altce of243 nrnc>ler/awzf?ollle (12.4 izzg for pla~nza-deritjed prod~tct\l 1 5 ursigned to the l" WHO Interrzatzrrnal Stcinrturd for AAT (05/1621 In line tz+ztlz the potency vulne.

As of 25 July 2008 l l of the 16 invited participants have responded. and all agreed with all of the proposals. In addition to voting on the proposals. parcicipanrs were also invited to comment on any aspect of the report. Four comments were made. the first was a general comment that the work was "cornpi-eheizsive irnd tt*ell done" with another stating "tve ugree tlzcit there i.r vai~re iiz assigning a protein and arztigerz tsalue to the ~tandurcl: These vulzies help t.~ithfiltur-e prr>dztct c*ompariso~zs ", In the third comment suggested "jte~foi-ming umir?a ucid urzu1y.ri.s orz ulE sunzples in order to hesr determine .~j?ec@ic activif?"' with which we agree and this will be done in due course.

A more detailed and specific comment was also made: ''My cont,errz with the rej7orr ;.S that rzo inforrnatiorz / cletail'r ut-e prol'ideci ccinc.ertzing the nzethnds icsrd hs the colEuhomting 1uhorutc~rie.s. What stantiurds tvef-e znsed,for the Bradford iissuys arzd antigenir. assuys;-' Were these ussays validated:.' Did afzyane rer2iei.l. the MW dafa?"

WNOfBS108.2092 Page 18

To address this concern: participants were asked to perform their in-house method for total protein Bradford assays and antigenic assays and analyse the data against their usual standard. and against a highly purified AAT standard which was provided. Details of the methods used were provided by most participants. typically citing a contmercial kit or standard. We haw presented the results here stating the type of the standard used without, disclosing further details. The reason for this approach is to determine the spread of results using the methods and standards currently utilised by different laboratories. Any inrjalid assays were determined using statistics and these results were excluded tkom the calculations. This is the traditional approach used by NIBSCIWHO niith establishing International Standards.

APPENDIX A:

CVHOfSS108.2092 Page 19

STABILITY OF IS (051162) AND CANDIIIATE PREPARATIONS A (05/150), B (051 152) AND D (051 172) In the first potency study report' the predicted long term stability of each of the candidate preparations A 10511501, B 10511 52). C (0511 621 and D (0511 72) was calculated by measufing the potency of accelerated degradation ampoules stored at 20, 37, and 35 "C from the filling date (a duration of 6-8 months at the time) relative to the potency of ampoules stored at -20 "C. At this time there was no significant loss of activity observed s~iggesting the preparations are very stable. Ampoules of each have been in continuous storage at -20.3, 20, 37. 35 and 56 "C and the stability of each candidate was re-calculated based on new data for 2.5 years. as described below.

Study design The stability study was carried out at NIBSC and invol\~ed measuring the potency of accelerated degradation ampoules of the IS (candidate C) (0511 62) and the other candidates A (05/150), B (051152) and D (051172) stored at -20, 20, 37, 45 and 56 "C for 2.5 years. The potency of each sample was determined by titration against elastase using the method described in Part 1 of this report. 12 separate assays were done in total. 3 for each of the 4 candidates with a11 storage temperatures measured in each assay.

Results and discussion The potency of the samples stored at 20, 37. 45 and 56 "C was calculated relative to the -20 "C sample. The relative potencies of these samples over time can be modelled to provide an estimate of annual loss of potency. Data was successfully fitted to the Arrhenius Equation and the predictions for $6 loss of activity per year for the IS (0511 62) and each ofthe candidates stored at -20 "C (normal storage conditions) is shown in the table below. The % loss per year is also shown for ampoules stored at +20 "C to assess stability during shipping at ambient temperature and the upper 95 ?E confidence interval of C/c loss at both temperatures is shown. Additional data are given at the end of this appendix.

l (upper 95 confldcnce hmlt) I (upper 95 confidence l l m ~ t i / IS (051162) 1 0.041 l 1.464

% loss of activity per year

Based on these results the predicted loss of potency for all of the samples under normal storage conditions 1-20 "C) is in the range O.Cf004 - 0.043 % per year. If the upper 95 9i confidence limit is taken to represent the worst possible case then the highest loss per year we could expect would range 0.005 - 0.375 5%.

-20 "C

The +20 'C sample is representative of storage at ambient temperature and the conditions used for shipping orders of the standard. Predicted loss of potency under these conditions is in the range 0.2 1 1 - 1.61 9 per year (0.004 % per day).

+20 OC

VCTHOI13S/08.2092 Page 20

Conclusion The results above indicate that the stability of the IS (051162) and each of the candidates from the potency study1 is acceptable under the current storage conditions and can continue to be skipped at ambient temperature.

Additional data:

A (051150) Storage

temperature "C - 150 -70 -20 +4

+20 +37 +45 +56

% loss per year

0.000 0.000 0.002 0.065 0.469 3.010

% Activity remaining ! Calculated rate Observed Predicted 1 constant

1 l

!

0.00000 0.00000 0.00002

I j 0.00065 94.30 1 98.84 l 0.00370 86.10 92.65 1 0.03057 90.20 1 84.19 60.40 61.10

WH01BSIO8.2092 Page 21

WHOII3S108.2092 Page 22

APPENDIX B:

LIST OF PARTICIPANTS IN THE LATEST STUDY TO ASSIGN POTENCIES TO WCOMBINANT PRODUCTS

Dr Cra~g Theluell Biotherapeutic\ Waemo\td\i\ Section NIBSC Blanche Lane South "tlimms Potter., Bar, HERTS EN6 3QC. UK

Dr Ewa Marszal Division of Hematology. FDAICBER, 29 Lincoln Drive. Bethesda. MD 20892

Dr Susanne Breitner-Ruddock Paul-Ehrlich-Institute FG 713 2. Paul-Ehrlich-Str. 5 1-59 D63225 Langen Germany

Alexander Stafford Manager Biophamaceutical Development Arriva Pharmaceuticals Inc. 1010 Atlantic Ave. Alameda. CA 94501

Dr Paul Bourdon CTC Biotherapeutics 5 Mountain Rd Framingham, hla. 0 1702 508-370-5373

WHOIBS108.2092 Page 23

APPENDIX C:

"M190 Bnirrtnnaboxai S w a r d 1st tntmaetirsnri SQn62fd f c i Mmr-IAntrt-#psm Pbsma.

Ben-& WBSC ccf* Of2362 fl*shuctons hi use

( V w i ~ w r Z G Gat& 2@%23;2:20081

2 CAU1%UN The prwaraaon 1% mot fur bwistrJBan to humans

Tl3e weprafiw ca-.ams m&.ia "r hJlrar wyir wll ewer m- iiral er 'he sawx mewaJi 'cm 21~21ptz3h it r, bc4M Pate M-

*estec and w,qd r;ril;rtue ':* f.Bb*a a-ir-41' arrrf WC$' RP& & W'*

? hlP11TSE Tl*e d me 1st rmtwqat mal ?m2u-r% ?W AA- vvas dffirmnmee a5 p"i d a c;lIahwcere SL* &em 'h = r r u b ~ , as&* tt., &A- vcas ix?mnfrd *at r-i ag+?st *W mdar arcsrwra:a;l d a% e *;g5Ir -5 aewmtree t=r &we rfe matm &-*a 4J+-rccPmp 4.

@"?win-* bmdiwc?r YSfl-ii ACf a m -G% - fan? a 'g@ ' l smc*ic#m% cwpe* W r w n r ni M T a g a ~ J k ? m cwmt&w d t-p-r atme ewessw at A4- ~ l n e m m BP d m @as -he m& at AAT - lrs-mitew S*a r sa4 IS 243 r m a e pw a-*% Thrs ii t3iltiatem b 1-4 o-q &W &CIF- *I

ampwe aFevlrg a ~ o i r c ~ a r G+ F * iW3 p e r m%te-*?m Q mss s ~ a - e m v iw -b S e ~ e z a a s r F w ~ o w m m ' m s Y& c;%e,e?t rnam&i.w weie-is at' -ME a 1-X miwscr *c caii;lia& Tr fq r m - 6 f-fs -ws r-tmr d me m m m a r Fww-arcr & & - s w d W&cr assays W*& #?a& Oh S'annsn ?twwt*mal 3,mrns m O & ~ ~ ~ : F S Y HI@-& Jra based 41

c - m s i ( f & r%fzess' 09 O ~ P $ Smn& "Ciq 3 pa*iaar mexswar-j a5 sum ire ?d a s a m W wc#air?y -W Lrcwarir* e t?s amem& M 35-1 -a# m mseerec as t% c n s F o m t o- z&na-mr ef Ice f: &c% was : l

F D3RECTiOM8 F M C ~ E M i f e 6 - -1 8 r n c ~ ~ p e ~ Pa~i'* m eSW- mI&"mz 5tm5s j3ci-ii Were xle ?a-C,) amuciae Set- jurna Clr a m p o d e mii;r -ac 'be e x L a gr+y m r red ?h --eqa rr, he ktm? rlatccc~ rrc; E-ts.rre : t i l ?'-r a t s c d > e a n ~ p A e sa'e:! xea+er vrwdeb ia $L%-d wan W @a stem ;i hii srxmiue s"J WITS ?W skouiaer a* ~IE

I T ~ ~ L - E bm'i *a0 De t d y n( rfie mca_ie 0 we %rri a ~ d :he drqxasabe wt~xie n-eaer caww m? am'w.e SW: be-*X. me t7bnls n o SE: Ji 9 e srhr -$am iipp? a e e 5 r g fr:e m wen h e a-33~'e a? tre *ntou& stress prr-al+ &sirVa b e barn ~elcnc; . - v* v-ast*: mf ar Cam s?;a;l oe ?Mm to ?As; rrc prq&Je ga5a Jraperrs ??a m;@ &er *m eym, era~~ie, oy IPn dse d g!je-+es arc! aq e$r staeb Take "a? re m j i e a l 15 imf $?m %& amp.& arc glass fa 3 ,r:o b e ier-ooue '&'h? ilf arwx8.a IS a y nrrcger jan at 5@ky ,esi mar atwa~'wli3 v e s s i ~ l l . A -a* ~ i m s s o + e tamc44~ steaker IS p m a W* r3tzI- Olsi m-&e

sea s-S-BWS a d mhii.3 r a 8- a5 a *wet% *"1~^cIes S%& m S%& a :2 C M km Zei are s i? imd 6 3 n G W%w?..ztms J%: arr u a b "or orimt m*&%) @&re bse Ite arqxa~ies shed3 me W-4 to

* q w m m me "ha m d tw ~ m l e ~bmd W - ~ m a ~ c M n i - J a ~ i 5 i l : & ~ & s ~ J w 1 r s ~ s r i l s r ~ m h ~ . ? r ~ C J ~

e s i a s ~ i t w F r w e w marrats are hea at Li'ESC &I a s s ~ d fenceraw+ mrruEe e ~ r s ~ -as:?re3 P m w w e Epiew.ii35 5 ~ 3 ~ 1 3 3e stared OP

'ecee ss rdr~rrl m h e l=@

12 CUSTWER FEEDBACK ,.t%tm-s are pncc&-i?@~C. :C pnmijp feea333k 3s- the slrit-v'?# x use ;I :W maim3 prcc-&& r mf ii zu' serdce ea%f :+M 3ny -=neqte :C srau-esalit.6; ac ~b

$2 crrAnt:tN 1- bf *3;iC+10'?5 .*>dCfT IS*& &"*?S, l" '&*G1 h- l W 8 i W a if

-e'went:e3, 4 5 mprv'an? va t 'be wwaratm's x, a - ~ t s h e :IIBEZ 33ce 1 m < m r W-3 Ice came am3 as*>% c* tdBSC a* CltSn a*CJ me3 ? a m ~ ; Q

14 WATERIAL %Af ETV SWEf3 i

h+ws.ca: and C k m t s a i

National lnstiruts for Biological SQanderd9 and Control

WHOIBSIO8.2092 Page 23

a BSC

16 UAEaL!M A M LOSS .Cw/,?car pc..curoed tq et* s g rim &W fie d A~

i r ~ s o ~ a O f ~ care and W:: ir 6 sml'at-6I. gr?wr&of3 a'?a .ss& h i s s m w t d wat~a* I;&.t)' to re Frtltipmn: n 4 s aqiii-~:lon ar* asE.

't 1 he WWS~%,&SIW nf L f i r ?~"$?lW3 '0 t"% a~W6Wi32W@EQ d :fe sDrctarc3 er reieran mxenais sdocbfa 2$ the t?3t+di+ ts :he leaprent The iZI- ihe m F a - O ap~~lrst io? mnb c?suIF. t PS h e c e c e s w ce&lizzd &I&? "c ~detw,le 3% * j x e agrpcc~nrnnsrr Fiestas% &";in& me Q-3ufr are ufioly b De 5-rc&alt c* condnu-ts of UK by Pt REIO@%~P: aqo Wie eana31kTt d mas-rms b q m c &e ceqt-ai d he hscfk-t+

A wa~~arBes rtre ~ x c i m e d to :?a i. e% ePmt ~3"~litted t-9 :zw. ,r"dmng ~ivcaf .wlzat-w hs. *P ~ 3 c a are fr@ iiw I P ~ ~ ~ O J L -:S cr eat t!w supohj of Chxs i r i l rat :?f-Pae r-jt-?S d ;ivy :h?d mfiy

The I?&+& %l-a:# ~3 k Ir*de m :Pe Rw~grert fw a-#, w""".c ass *=@er CC-W: W *r,5ir?f-i, wn'm ansr r ow%w2i;r .wtir, W 5

%Fee-mi

Tw t-i iianl1:y -3m5 if3stt~ffe n CC-T~CXIO~ &q h i s ai;-@r&m, W"EL3er hf mslE+l7r3E W WEX?? Of C5WdC? DT l?t?%-?VIW 4"ai. k M

c * , - c d -;Z% d s-r! 3nce p a d of Gayawe iirt fm R?CI'P@: S*& Of IF* sm3s

l* a r r d w?r &z6-F %ld/3ekd 3~ l"+. ~ e % & e m~.+Ib ~ o v e m? L3 F a

t -w- we9catcn ~d-en %tor& am L-J o o ~ % & ~ :a-3 w-~?.ided eat the Re-ZD& Pas 7 e t ~ w i tee ZGmc-, ?c tw 1,75tititiCe f-:~ge?hW .&:P

W&P f~-fiiCJsm C* SUS- g l i ~ e c rxflrm wsh~n s e e n ha" c't"+~ We

PJilrionai lrtsrirute lot Biulogieai Stsrtdards and Cortaroi

Page 2 of 2

?VHOfBS/O8,2092 Page 25

APPENDIX L):

EXPERT GOMLLIITTEE ON BIOLOGICAL STANDARI)ISATION GENEYA 23-27 OCTOBER, 2006 ENGLISH ONLY

INTERIVATIONAL COLLABORATIVE STUDY TO ESTABLISH THE \TWO lST INTERNATIONAL STzANDARD FOR ALPHA-l-ilNTITRYPSIN. (0511621

Craig ~he lwel l ' . Peter ~ i ~ s b ! ' and Colin ~ o n ~ s t a f f ' '

I Biorhet-uf)erttic.s Grorlp, Nne,nosta.si,~ Section and '~iit.statistic*.s Section, Nutintzal It~sritltte for Biolrtgicul Standards and Corztrol, Sorlfh 'Mimms, Herts EM6 SQG, U K

\.VHOnBSfO8.2092 Page 26

An international collaborative study was organised to establish the I 'I International Standard !IS) for Alpha- l -Antitrypsin IAAT) as agreed at the Alpha- l Foundation Workshop. April 2005. Cincinnati, USA. The study involved 15 laboratories from 10 difi'erenr countries. 1,aboratories were provided with detailed methods and critical reagents and were asked to measure the potency of3 candidate standards (A. 051150: B, 05/15?: C. 051162 and D, 051172i. In addition laboratories with relevant experience were invited to carry out extended characterisation of the materials. including total protein and antigen content for which an additional reference preparation was provided. Analysis of the data indicated that any of the candidates would be suitable based on potency determination. Initial stability studies foilowing 6 - 8 months of accelerated degradation at elevated temperatures showed no loss of potency for any of the candidates. Candidate C had the lowest variability of filling. and the largest number of ampoules available, it is therefore proposed that candidate C (0511 62) be adopted as the I " International Standard for Alpha- 1 -Antitrypsin with a potency of 243 nmoles ( 1 2.4 mg) active AAT per ampoule. The remaining candidates could serve as replacement standards (in the order A. B and D) once C has been depleted, dependent on the stability of each material. Total protein and antigen content was determined for each candidate which also allows for calculation of specific activity. The results were variable between laboratories and between the different methods and depended on the protein standard used. Further complications arose relating to the molecular weight cited for the AAT standard, where this was used. Additional discussions or studies will be needed before a value for total protein or antigen can be assigned. Therefore, at this stage we do not recommend assigning values for total protein, antigen or specific activity to the AAT International Standard.

INTRODUCTION

AAT belongs to the serpin (serine proteinase inhibitor) family, however its physiological target is elastase. Plasma derived therapeutic AAT products are used to treat Alpha-] -Antitrypsin Deficiency (Alpha-l), a genetic disorder identified in virtually all populations that can cause liver and lung disease in adults and children. Alpha-1 is often misdiagrtosed as asthma or smoking-related COPD (chronic obstructive pulmonary disease), and as many as 3 C/c of all peopie diagnosed with GOPD have Alpha- 1. The World Health Organization (WHO) American Thoracic Society (ATS) and European Respiratory Society (ERS) recommend that all individuals with COPD be tested for Alpha-I.

A need for reference preparations to standardise the measurement of potencies of therapeutic products of AAT has been identified by regulators and industry. The European Official Control Authority Batch Release (OCABR) guidelines for 2005 list soluhilitt; urtd upprarur?c-r and potenq as requirements, and a European Pharmacopoeia monograph is currently being prepared by Group 6B. which will depend on the establishment of an IS for potency deteminations.

In the absence of an IS and therefore an International Unit (IU) for AAT, and in line with the current drive to extend the use of SZ units where possible. an agreement was reached to calibrate a new IS in terms of moles (andlor milligrams) of active inhibitor per ampoule. This is also in line with the current labelling practice of licensed AAT therapeutic products.

A plan for this study, including the choice of materials and methods, was agreed by representatives from academia. industry and regulatoq authorities (including NIBSC. FDAICBER and Group 6B of the European Phamacopoeia) at the Alpha- l Foundation

WHOn&S/O8.2092 Page 27

'rh,Iorkshop. April 2005. Cincinnati. USA. The plan was finalised at this meeting by a sub-group of participants (Appendix A). including major manufacturers of therapeutic products.

Intended use of the standard

Although the IS will be defined in S1 units based on absolute measurements using the methods in this study, the same rnethods are not intended to be used to assign potencies to unknoaln preparations. Rather the IS will be used as a reference preparation in comparative assays, typically titration against the physiological target elastase, to assign potencies to test samples relative to the IS in mofes or milligrams.

The potency of future replacement standards could either be assigned by absolute measurement of active inhibitor. as with this study. or based on the potency of the previous standard in a traditional comparative study.

PARTICIPANTS

A total of 17 of the 21 invited laboratories from 10 countries agreed to participate in the study. 15 laboratories returned results (US (5 ) . France (2), UK. Germany. Spain, Ireland, Austria, Su~eden. Israel and Australia) including 5 regulatory laboratories. 1 academic and 9 from industry. Participating laboratories are listed in Appendix B. Each laboratory was assigned a code number and the order of listing does not correspond to the identifying number assigned to each laboratory.

MATERIALS

Candidate niaterials were provided by four manufacturers: Baxter Bioscience, Vienna. Austria; ZLB Behring, Illinois, USA; Talecris Biotherapeutics, North Carolina. USA: and Arriva Pharmaceuticals Inc, California. USA. Baxter. ZLB Behring and Talecris provide most of the therapeutic AAT used worldwide at the current time. Several smaller scale trial fills were performed to optimise freeze drying protocols. For the definitive fills. all manufacturers were requested to provide 100 g of material sufficient for a fill of 10000 ampoules at 10 mg per ampoule. Baxter Bioscience provided 200 vials of 0.5 mg freeze dried material: Talecris Biotherapeutics provided 2 litres of material at a concentration of 50 - 55 mg/ml; ZLB Behring provided l00 vials of 1 g freeze dried material. Each of these was prepared from pooled human plasma that has been tested and found negative for HBsAg, HIV antibody, HCV antibody and HCV RNA by PCR. Arriva Pharmaceuticals provided 200 m1 frozen material at a concentration of 50 mglml. This material was recombinant. expressed and purified from yeast. Manufacturers provided package inserts or other details of fractionation and production where appropriate and this information is stored with the relwant filling records.

The four batches of material were reconstituted andlor diluted to a final concentration of 10 mglml A'AT (based on the manufacturer's information) in 17 mM phosphate: 38 mM chloride, 81 mM sodium at pH 7.0, hiiannitol ( 144 mM) was added as a bulking agent to a11 materials except candidate B, where albumin naturally present from the purification process served this purpose. AmpouIes were prepared conesponding to the 3 AAT preparations and were coded 051150: 0511 52: 051162 and 0511 72. and designated for the study as candidate preparations h, B, C and D respectively (note: this does not correspond to the order of the manufacturers listed above). The filling and Iyophilization of the candidate materials was completed successfully,

WHOISS/O8.2092 Page 28

following WHO procedures. The a~npoules were filled by and are stored at NIBSC. Potters Bar. Ir'K, at -20 "C. in the custody of Dr C 1,ongstaffand Dr C Thel\vell. Detailed physicochemical characterisation of the filled candidate materials was performed as otitlined below (see exwfzcled c.l?urzrc~rrvr.isution 1.

For participants carrying out total or antigenic protein concentration determination a reference AAT preparation was made available from Athens Research and Technology (ART), Athens, Georgia, USA. This AAT preparation (lot no. PI2006-WHO; catalogue no. 16-16-01 1609) was prepared from human plasma by a non-commercial-scale. chromatography procedure. A protein concentration of 19.2 yM was assigned to the preparation by ART, based on an A280 value of 0.449 applying an E"' of 0.433 and a molecular weight of 54 kDa. The assigned concentration based on functional activity was 18.7 yM. Purity was stated as greater than 99 % by SDS-PAGE under reducing and non-reducing conditions. The material was prepared from pIasma shown to be non-reactive for HBsA. anti-HCV, anti-HBv, and negative for anti HIV 1 & 2 by FDA required tests.

Stability of candidate reference preparations

Each of the candidates A (0511 50) B (0511 52) C (,05/162) and D (05/172) were subjected to accelerated degradation trials by storing ampoules at -20. 20. 37, and 45 "C from the filling date (see table above). The potency of each of the candidates at the different storage temperatures was measured by titration against trypsin, using the pIate reader method described in this report. Four laboratories (NIBSC. Potters Bar, UK; Baxter Bioscience Bioscience, Vienna, Austria; Octapharma PPC, Vienna, Austria & Paul-Ehrlich-Institute. Langen. Germanyj were asked to perform four separate assays. one for each of the candidates, including triplicate measurements of four dilutions of each storage temperature. If time was pemitting the laboratories were asked to repeat this with fresh ampoules and sample rotation within the plate. All four laboratories returned results, with three also carrying out the requested repeat. The potencies of the sanlples stored at elevated temperatures were compared to those stored at -20 'C and the results expressed as relative %h activity are provided below.

FVH01BSf08.2092 Page 29

/ (7 months) 1 (97.21 - 110.55) I 195.00 - 103.98) 1 193.96 - 108.36) / C I 107.00 , 99.78 i 101.06

From the results above no significant loss of activity was observed over the stated time period. Ongoing accelerated degradation studies will determine the long term stability of the materials. Ampoules of each of the candidates are in continuous storage at -20,4, 20. 37.45 and 56 "C and will be assayed periodically. The nature of this study offers an advantage for monitoring stability. as it is possible to measure the absolute potency of the -20 "C samples using the methods described below. Traditionally the degradation of samples stored at elevated temperatures are measured by comparison to a reference temperature (often -20 "C) and assume that the reference sample is stable. Because we can measure absolute potencies against active- site titrated trypsin, we also have the option of measuring absolute loss of activity at all temperatures.

An additional indication of the long term stability of AAT preparations is provided by earlier stability data of a trial fill from a previous investigation involving two plasma-derived. and one recombinant AAT product, in 1999. A number of ampoules of each of these materials were stored under accelerated degradation conditions at 4. 20. 37 and 45 "C over a period of 4 years. Assays show that AAT standards retain 75 % of their original activity even after storage at 45 "C for 4 years. The relative potencies of these samples over time can be modelled to provide an estimate of annual loss of potency. Data was successfully fitted to the Arrhenius Equation predicting a loss as low as 0.2 Cic per year at -20 "C.

In all our studies so far we have no indications of problems with stability of samples kept on ice after reconstitution. End users will be advised that if they wish to store samples for extended periods after reconstitution. or freeze and thaw aliquots. they should evaluate these procedures themselves.

ASSAY METHODS AND STUDY DESIGN

Assigning a potency value to the candidate reference preparations

To assign a potency value to a preparation of AA1' a titrant of known activity must be used. In the absence of an active-site titrant for elastase, the physiological target fbr AAT, porcine trypsin was substituted. AAT inhibits trypsin in the same way as elastase and it was agreed that porcine trypsin would be the most suitable replacement based on the recommendations of manufacturers and other experienced groups. The first part of this method is therefore to determine the 5% activiry of a batch of trypsin. lbllowed by titration of AAT against this trypsin.

WHOR3SIO8.2092 Page 30

For the first part of the study participants with experience of active-site titrations were asked to titrate trypsin against 4-Nitrophenyl 3-guanidinobenzoare hydrochIoride (NPGB). Csing this method, the s~thstrate N K B is added to a solution of tvpsin causing a burst of activity and retease of chromophore, but the substrate is turned over \erg; slowly. The magnitude of the burst. quantified by reference ro a standard curve of 4-nitrophenol. is equivaleilr to the molar concentration of active trypsin. Each panicipant was provided with a detailed protocol and the following reagents supplied by Sigma: NPCB (Catalogue no. 51 05: Lot 42658311). Type IX Trvpsin from Porcine Pancreas (Catafogue no. T0303: Lot 045K7775) and 3-Nitrophenol. PESTANAL. analytical stalldard (Catalogue no. 35836: 1,ot 3278X).

Panicipants were requested to provide accurate details of the trypsin solution prepared to determine the total trypsin concentration. This included: the amount of trypsin weighed out and absorbance readings at 280 ntn (A280) of dilutions in triplicate. The trypsin concentrations were calculated using a molecular weight of 23.3 kDa and an extinction coet'ficient (E'%) of 15.0 as stated on the product information slieet provided by Sigma. Five different dilutions of trypsin were titrated against NPGB with the absorbance at 402 nm monitored over 3 minutes. Three replicate assays were requested, v~iith the raw data to be returned to NIBSC for analysis. In addition participants \\>ere requested to generate a standard curve of 4-nitrophenol (using the analytical standard preparation provided) in triplicate according to the protocol provided.

Titratiorz of it4 T against trypsin

All participants were asked to perform this part of the study. A known molar concentration of trypsin (determined above) is mixed with increasing amounts of AAT. Since trypsin and AAT are known to form a 1: l complex, extrapolation to 100 % inhibition of the AAT titration curve can be used to calculate the molar concentration of rlAT present. Each lab was provided with a detailed protocol with methods for kinetic and endpoint assays in 96 well plate or cuvette format. Only one of these methods was required, with the choice made based on the equipment and experience available. In addition each lab was provided with the following reagents supplied by Sigma: Type IX Trypsin from Porcine Pancreas (Catalogue no. 1-0303: Lot 045K7775j. N,- Benzoyl-L-arginine 3-nitroanilide Hydrochloride (L-BAPNA) (Catalogue no. B3 133; Lot 095Kl482) and Alpha-l -antitrypsin candidates (6x 10 mg (approx.) ampoules of the four candidates: A. B. C and D).

Participants were again requested to provide accurate details of the trypsin solution prepared to determine the total trypsin concentration. This included: the amount of trypsin weighed out and triplicate A280 readings. In each assay, complexes were formed between dilutions of A.4T and a fixed concentration of trypsin. The residual trypsin acti\fity was nleasured using the chromnpenic substrate L-BAPR'A. For the plate reader method 4 different dilutions of each AritT candidate were used in triplicate, with 4 assays requested. In each of the 3 assays the position of each of the candidates kvithin the plate was rotated to prevent any positional bias. For the spectrophotometer method 2 assays of 5 AAT dilutions were requested. Blanks urithout trypsin and trypsin alone (no AAT) were included in both methods. Participants were requested to return raw data to XIBSC for analysis.

Total protein concentration determination by Bradford assay.

Paaicipants were invited to determine the total protein concentration per ampoule of each candidate by the Bradford assay. to allow the calculation of the specific activity of AAT, KO

WHOfBSIO8.2092 Page 31

protocol was provided and participants were requested to use the standard protoocl utilised in their laboratory using their usual in-house reference standard. in addition to the reference AhT preparation from ART.

Antigenie determination of AAT concentration.

This part of the study was designed to determine the feasibility of utilising the candidate reference materials as an antigenic reference standard. as well as a fuunctional activity standard for AAT. No protocol was provided and participants were requested to use their usual protocol and in-house reference standard in addition to the reference AAT preparation from ART.

Extended characterisation

Laboratories with expertise in characterising AAT using a range of techniques were invited to provide additional analysis of each of the candidate materials. Data was invited from SDS- PAGE. Isoelectric foc~~ssing IIEF). mass spectrometry (MS) and size exclusion chromatography (SEC).

RESULTS

Assigning a potency value to the candidate reference preparations

Raw data were returned by 12 laboratories. A molar extinction coefficient for 4-nitrophenol was calculated from a standard curve for each lab. using a molecular weight of 139.1 1 (Table I ). Active-site titration data were analysed by plotting absorbance at 402 nm against time and fitting a curve describing a burst followed by steady state phase. The linear steady state phase was extrapolated back to time zero and this absorbance corresponds to the initial burst of 4- nitrophenol. The moles of 4-nitrophenol released in each burst was calculated using the molar extinction coefficient for each lab. and also the mean of all the 4-nitrophenol extinction coefficients from the study (an overview of the analysis of the active-site titration data is given in appendix C}. These values correspond to the moles of active trypsin for each dilution. The mean of these values was taken (corrected for the dilution factor) and an overall mean and standard deviation was calculated from the three replicate assays from each lab. One-way analysis of variance was carried out for the data from all labs for each trypsin dilution. and there was no significant diffirence between the results obtained for each dilution.

The total trypsin concentration was calculated for each lab based on both the weight of trypsin used. and on the A280 values provided. The mean active trypsin based on individual 1aboratot-y"~ molar extinction coefficient for 4-nitrophenol was expressed as a percentage of the total trypsin for each lab by weight (Table 2: Figure 1 ) or by A280 (Table 3: Figure 2). Son~ervhat surprisingly the variability amongst A280 deteminations of trypsin concentration was rather high. The mean active trypsin based on the mean molar extinction coefficient for 4- nitrophenrtl from all laboratories' results was expressed as a percentage of the total trypsin for each lab by weight (Table 3: Figure 3) . The column chart in Figure 4 shows the distribution of data from all laboratorjes for each of the methods used for measuring total trypsin. and extinction coefficient.

'ClrHOIBS/08.2092 Page 32

A value of 82.3 % active trypsin was returned when the total t rysin was calculated based on weight using either the indit~idual laboratory's molar extinction coefficients for 4-nitrophenol. or the overall mean fionl the srudy. with CVs of 5.5 and 4.2 Q respectively. A value of 85.1 9; was returned when total trypsin concentration was calculated from the A280 values with a CV of 22.2 72. reflecting the variability in concentration measurements determined by 21380. An active trypsin value of 82.3 ";: %vs therefore used to calculate the potency of the .4AT preparations in the next part of the stud). based on trypsin stated protein content by weight.

The results from laboratories 10 and 14 were excluded from the final analysis as statistical outliers, and are not shown. For laboratory 10 the active trypsin calculated tvas higher than the total trypsin value deternined. This was most likely due to the unusually low extinction coefficient calculated from the standard curve of 4-nitrophenol, though discussions with the laboratory failed to identify the source of this anomaly. For laboratory 14 the active tt-ypsin calculated was approx. two-fold lower than the mean of all other laboratories in all instances. 4 likely cause for this is an error in the preparation of the trypsin solution. The results from laboratorq 12 were returned after the trypsin titration analysis was complete, and could therefore not be included for this part of the study.

Data was returned by 14 participants in the form of endpoint values (3j, calculated rates (2) or raw data files containing absorbance values against time (9). Where necessary rates were calculated from the raw data; and for each assay the rate or endpoint was plotted against volume of AAT in the assay. The volume of AAT required to completely inactivate the trypsin was determined by linear regression and extrapolation to the x axis,

The total trypsin concentlxtion was calculated based on the weight of trypsin rather than the A280 values provided. This decision was based on the results of the active-site titration where CVs were much lower when the weight was used. The active trypsin concentration was taken to be 82.3 C/c of the total based on the active-site titration data above. With the active molar trypsin concentration for each assay known, the molar concentration of active AAT was calculated. The results for each laboratory for candidate A are shown in Table 5 and Figure 5. The results for candidate B are shown in Table 6 and Figure 6. The results for candidate C are shown in Table 7 and Figure 7. The results for candidate D are shown in Table 8 and Figure 8. Details of determining assay validity and statistical outliers are provided below. The distribution of the results for all laboratories is shown in Figure 9 for candidate A: Figure 10 for candidate B; Figure 1 1 for candidate C and Figure 12 for candidate D.

A summary of the mean potencies for each candidate calculated with statistical outliers removed is given in Table 9. The results are given as both nmoles and milligrams of active AAT per ampoule. These nmole results are converted to milligrams based on a molecular weight of 51 kDa for plasma-derived material and 44.5 kDa for recombinant APIT. These molecular weights are based on LC/MS data provided by one of the manufacturers as part of an extended characterisation of all of the candidates after filling.

Det.iution.rfrom the pro focctl

Each laboratory performed the assays according to the protocol provided with the following exception: laboratory 13 substituted the tris working buffer (0.2 M Tris-HCI, pH 8.0) with TNB buffer (0.05 M Tris-WG1; 0.05 M NaCI: 0. l 66 BRIJ: pH 8.0).

WHOIBSIO8.2092 Page 33

In each ascdq and for each cand~date rewlti here excluded uhere lrnear regrc\cton of acticlty venuc AAT Miar not fea\rbIe

* The re\ult\ from laborator) 12 for cand~date A. d \ \ d ~ 2 and candidate B. Jiiaq 1 mere excluded I-rorn the analycic In both c;t\ec the c(- actit it) for louect AAT concentration t%ac higher than that for the tryp\rn &lone (no AAT).

* 717e reiult5 from laboratory 13 for all d \ \ q s of candldate5 A. B and L) were excluded from the andlyw,. In all ca\ei the value for luwe\t AA7 concentration mac h~gher than that tor the tr)p\in alone (no AAT)

Results of faboratoy means that were greater than 2 standard deviations from the overall mean for each candidate were considered statistical outliers and excluded from the final analysis:

* The results from laboratory 3 for candidates A. C and D were excluded from the final analysis. Each of these results was greater than 2 standard deviations (SDI below the overall mean. For sample B the results from laboratory 3 fell just within the 2 SD limit. however there appears to be a problem with the method, most likely substrate depletion which would underestimate the amount of AAT. It was therefore decided to exclude the results from laboratory 4 for methodological reasons.

While the results from laboratory 3. using endpoint values with the plate method, were excluded on methodological grounds; overall the results from all of the other methods used in the study showed no statistical difference.

Several sources of uncertainty in calculating the potency of the candidate materials may be considered. The largest source of variation in the calculation of AAT potency was identified as the variation between laboratory mean potencies. The uncertainty of measurement for each AAT is expressed as the 95 % confidence level of the interlaboratory means (table 9).

Another possible source of variation comes from the measurement of active trypsin, from which the AAT potencies are derived. This is dependent upon accurately measuring the amount of 4- nitrophenol released from NPCB by trypsin. The absorbance from the active-site titration is converted to moles of 4-nitrophenol using an extinction coefficient. Laboratories calculated an extinction coefficient for 3-nitrophenol using an analytical standard IPESTz4NilL). This reference preparation was the highest grade available commercially following the discontinuation of the National Institute of Standards and Technology (NIST) standard SRM938. NIST were approached to provide material for this study. but they declined. Va~ation introduced by measuring the extinction coefficient of the 3-nitrophenol analytical standard ultimately provided was minimised by using a mean value from all of the laboratories' data. Overall variation in the extinction coefficient was IOW with a CV of 5.20 (Table l ,I and the mean 5% active trypsin calculated fiom individual laboratories. results had a CV of 5.88 56: and when the mean extinction coeffit'lent was used the CV was only 3.45 %.

WWOfBS/O8.2O92 Page 34

Variation in the filling is also a source of uncerlainty of measurement cited for Inremational Standards and was low in the study with CV values ranging between U. 15-0.23 Ci for the 3 candidate preparations (see above).

Total protein concentration deternrlnation by Bradford assay.

Six taboratories returned results for totaI protein detemination: 5 results using the ART reference preparation as standard, 6 with an in-house reference preparation (AAT 12) albumin 13) Semm 11 1). The results using the ART standard are shown in Table 10. The results are given as r~lilligrams of protein. !?;here results were given in terms of nmoles, these were converted to milligrams using the stated molecular weight of the ART AAT reference preparation of 53 kDa. The results using the laboratories' own in house standards are given in Table I I . Where these results were given in terms of nmoles, the conversion to n~illigrams was done using the molecular weight provided by the individual laboratoq fbr their own standard.

Antigenie deternnination of AAT concentration.

Six laboratories returned results for the antigenic detemination: 5 results using the ART reference preparation as standard, and 5 with an in-house reference preparation (AAT (2). and serum (3)). The results using the ART reference preparation as standard are given in Table 12. Where results were given in terms of nmoles, these were convened to milligra~ns using the stated molecular weight of the ART AAT reference preparation of' 54 kDa. The results using the laboratories' own in house standards are given in Table 13. Where these results were given in terms of nmoles, the conversion to milligrams was done using the molecular weight provided by the individual laboratory for their own standard.

Extended characterisation

The molecular weights of each of the candidates by MALDI MS was as expected based on the theoretical sequence of around 50 kDa. Sample D had a molecular weight of around 6 kDa less, consistent with the absence of glycosylation in a recombinant product. In addition Electrospray MS can distinguish between the different isoforms, and showed sarnples B and C to have an almost identical isofhrm pattern, with A shifted approximately 150 Da though the reason for this was not investigated further. Again D showed a completely different pattern confirming the absence of glycosylatiun.

Native and desialylated isoelectrofocussing (IEF) was performed 011 the 4 candidate materials confirming the absence of glycans in sample D. however 2 isoforms were evident. The profile samples of A. B and G were comparable showing two major intense bands conesponding to the common M4 and M6 AAT isoforms. Samples A and B had more of the less abundant isoforms M7 and M8.

SEC-HPLC was used to assess the level of aggregation in the candidate materials. The lerrels of monomer, dimeritrimer and aggregates were comparable in A. G and D (monomer >90 % i c ) ,

whereas B had significantly more multimers (monomer 77.3 %?.

SDS-PAGE (reduced and non-reduced) confirmed the lsrmilar rt~olecular ue~ght i tor A. B and G, and 5howed addrttonal bandr for B. O~eral l A and C 5houed the hrghert punt?.

RTHOBS/08.2092 Page 35

CONCLUSIONS

The results of the potency tests indicate that each AAT candidate is potentially suitable to serve as the International Standard (IS) for AAT potency. The extended characterisation confirmed the similarity between the plasma-derived material of candidates A. B and G, and confirmed the expected differences in glycosylation of the recombinant product D. No clear prekrable candidate emerged from the collaborative study, so the choice of proposed IS was made on practical grounds based on the results from filling and extended characterisation. Candidates A and C had the lowest CVs in the potency determination, and showed the highest purity in the extended characterisation. however candidate C had the lowest variability of filling isee Materials section above) and had the largest number of ampoules available. We therefore propose candidate C (051162) as the 1" IS. From the results of this study, candidate C contains 233 nmoles (12.4 mg) of active AAT and we propose to include both of these values on the label.

Because of the overall suitability of each of the candidates, the remaining candidates could serve as replacement standards once the first standard has been used up, provided the material is stable and subject to further approval. Based on the same criteria as above, the candidates could be used in the order A (0511 50), B (0511 52) and D (0511 72) with the potencies listed in Table 9.

The specific activities based on the antigen and total protein tests provide an indication of the proportion of active protein in the ampoules of each of the candidates. Protein concentrations assigned using the ART reference material were calculated based on a molecular weight of 54 kDa, as stated on the ART data sheet. which is a commonly quoted molecular weight calculated from SDS PAGE. The potencies assigned to the candidates in this study were based on molecular weights of 51 kDa (plasma) and 43.5 kDa (recombinant) based on LClMS data. Calculation of antigen and total protein rnight require more discussion and consensus. or further studies, before a final agreement could be reached. The results between laboratories, and the differences between the different methods, are quite variable and we therefore propose not to assign a total protein. antigen or specific activity value to the standard at this stage.

Participants' response

A positive response to our proposals was particularly important in this study where there were many manufacturers with existing products, or products in development, and regulators, who would validate the proposed Standard as fit for purpose. Each of the participating laboratories was asked if they agreed or disagreed with the following proposals:

I . Mre 17ro11ore canclicfute C (0Y162) u.5 the I" Intert-~atioizul Stanllurdfor Alphu-I- Antitrypsin, and u.s.si<;.n u potcvzcj tlultte ($243 nt7zole.s actit*e Aiif T per ampoule.

2. This pofenc~. sh~lild he conl.erted to 12.4 mg iic,rz\*e AAT hu.ied on a nzoleclllur vveiglzt o f 51 XDcr (deten~zined h\- LCAlilSI and be itzcluded on tlze iufiel.

3. Pro~~ided the candidates are srahle, tlzr relilacernent Intert?utioncrl Starzdurdr. for rlAT coilld be c~undi&te.s A flW1501; B (05/152) uad D (fi5/1721, in tlzur order: lrsiny the pote/zcies derit~ed in this stltds.

As of 4 July 2006 14 of the 15 invited participants have responded. A11 participants agreed with proposals I and 2. and 1 1 of these also agreed with proposal 3. The comments raised regarding the replacement of candidate C were largely concerned with the suitability of the recombinant

Vf7NOfSS/08.2092 Page 36

material (candidate D) for this purpose. All agreed that candidates A and B were suitable as rrplacement standards.

Part~cipant\ \%ere a h o invited to comment on the following jtaternent.

111 the report u3e prrjprjst. tzot fr i ussig~? u .ipecifi(. ucri\-i& to rlze ccrndiduirvs 1iast.d otz the rex111ts of' this sttidy. EVtj uL.so ruisc; the prtssibili~ c!fussigrting a .spt.c.(fi'c ucti13ity r c ~ tize I"' 1.7 in ri?~~fiititre, possibly by u r , u r ~ i n g out a larger sturfs. Plt'ase jlro-i,vidf> any rhorlghts or r50rttnrents ?*oil 1iur.c. o r z this issur li.~lotr. Five laboratories pro~ided comments on this issue, with all agreeing that a larger study would be required. with two laboratories volunteering to participate in a larger study. Several laboratories commented upon the different results obtained for total protein and antigen determination. For total protein determination one lab suggested that calibration against kjeldahl would be needed. another lab suggested using UV absorbance with extinction coefficients determined by differential refractometry, but agreed further discussion would be required. For the antigenic determination one lab suggested that because of the high purity of candidate C, this could be assigned a 100 54 antigenic concentration. and the antigenic concentration of other preparations could be determined using this standard candidate.

Other general comments that were made included one lab also suggesting that establishment of a plasma standard for determination of AAT activity in whole plasma or serum would be useful. Another lab commented on the benefits to patients knowing that the potency of one product is comparable to another.

ACKNOWLEDGEMENTS

We are extremely grateful to the following individuals and organisations for their invaluable contribution to this study:

Symma Finn for organisational help, and the Alpha-] organisation for sponsoring the Alpha-l Foundation Workshop. April 2005, Cincinnati, USA; and to Prof. Mark Brantly and Prof. Geraid McE;lvaney for co-ordination of discussions.

Plttendees of the standardisation session of the Alpha-] Foundation Workshop. April 2005, Cincinnati. USA and to the standardisation study steering group (Appendix A) for helpful discussions.

The members of the project team from the Centre for Biological Reference Materials at NIBSC, Michelle Anderson, Paul Matejrschuk, Patrick Duff>/ and Sharon Coughlan. for development work on filling and organisation of sample shipping.

Baxter Bioscience, Vienna. Austria: ZLB Behring, Illinois, USA; Talecris Biotherapeutics, North Carolina. USA: and ,4rriva Pharmaceuticals Inc, California, USA. for kindly donating the material for this study. for sharing methodologies, data and for useful discussions.

Dr Andrew Shrake and Dr ER-a Marlizal (FDAICBER) for their contributions to the planning stage of the study. and for arranging the distribution of samples and critical reagents ( ~ ~ i t h Christine ctndersonf to participants ~ f i th in the USA.

WHOIBSIO8.2092 Page 37

Dr Virginia B. Pate1 and Athens Research 8r Technology Inc for supplying and distribttting the reference standard for the antigen determination. and to the Alpha-l Foundation for funding this part of the study.

All participants who took part in the study, listed in Appendix B. arid to everyone involved who is not identified personally.

WHOIBSf08.2092 Page 38

Table l . Molar extinction coeW~cient values for 4-nitrophenol. Results calculated from a standard curve for each laboratory using a molecular weight of 139.1 I . The mean. deviations and coefficient of variation (CV C,;-) are given. The result from laboratory 10 was excluded from the analysis as a statistical outlier.

Table 2. Active trypsin (%) by weight based on individual laboratories' results. Calculations using the mean active trypsin from three assays for each laboratory. Results are based on the individual laboratory's molar extinction coefficient for 4-nitrophenol and total trypsin calculated by weight.

WHOlt3SfOS.2092 Page 39

Figure 1. Active trypsin (Q) by weight based on individual laboratories+results.

Trypsin activity by weight (individual laboratory's molar extinction coefficient for 4.

nitrophenol) 100

95

Mean

- - - - - - - - Confidence limits (95 %)

50

0 1 2 3 4 5 6 7 8 9

Lab number

Table 3. Active trypsin (%) by A280 based on individual laboratories' results. Active trypsin (%) calculated using the mean active trypsin from three assays for each laboratory. Results are based on the individual laboratory's molar extinction coefficient for 4-nitrophenol and total trypsin calculated by A280.

Laboratory number

f - - * I n o \

l Total ve -- 1 trypm

I 1 1 1 1 1 Inmole\)

I / 165.87 i 202.39

Active trq pstn

81.96

Intra- laboratory standard de\ tatton 0.33

Mean Standard deviation

CV %

WHOIBSI08.2092 Page 40

Figure 2. Active trgpsin I%) by A280 based on individual laboratories~esults.

Trypsin activity by A280 (individual laboratory's molar extinction coefficient for 4-nitrophenol)

130

Mean

- - - - - - - - Confidence limits (95 %)

Lab number

Table 4. Active trypsin (9%) by weight based on the mean molar extinction coefficient for 4- nitrophenol. Active trypsin (5%) calculated using the mean active trypsin from three assays for each laboratory. Results are based on the mean molar extinction coefficient for 4-nitrophenol from all labs, and total trypsin calculated by weight.

li'HOIBSIO8.2092 Page 41

Figure 3. Active trypsin (%) by weight based on the mean molar extinction coeMScient for 4-nitrophenoi.

Trypsin activity by weight (mean molar extinction coefficient for 4-nitrophenol)

120

Mean

- - - - - - - - Confidence limits (95 %)

50 0 1 2 3 4 5 6 7 8 9

Lab number

?VHOR3S/O8.2092 Page 42

Figure 4. Distribution of data from the active site titration of trypsin based on different methods of analysis. Active tr>fpsin was caicrtlated based on mean or individual Itiboratofies' molar extinction coeffjcient values for 4-nitrophenol. Total trypsin was calculated by either weight or by A280. Colurllrts reprpsrnt the number of laboratories with results in the comesponding active trypsin range, and each box represents an i~ldividual laboratory by number. The mean value of 62.3 5:. based on both weight of trypsin for individual Iaboratories and the mean extinction coefficient for 4-nirrophenol. is highlighted on the chart.

WHOIBSIOS.2092 Page 43

Active trypsin (%) by active site titration

Mean

Weight (individual extinctinction 0 coefficient of enitrqlhenol)

A280 (individual extinctinction coefficient of Cnitrophenol)

Weight - (mean extinction coefficient of 4-nitrophenol)

50. 60. 70. 80. 90. 100. 110. 120. 130.

Trypsin activity (%)

lVHOlfZSf08.2092 Page 34

Table 9. Mean potencies of candidate A (nmoles AAT per ampoule). Potenctei are calc~tldted from the mean re%&\ from a11 the \altd d%idly\r returned.

Figure 5. Mean potencies of candidate A (nmoles AAT per ampoule).

AAT potency of sample A (0511 50)

Mean

Confidence - - - - - - - - limits (95 %)

0

0 1 2 3 4 5 6 7 8 9 1 C 1 : 1 2 1 3

Lab number

WHOIBSf08.2092 Page 45

Table 6. Mean potencies of candidate B (nmoles AAT per ampoulel. Potenciw are calculated from the mean rei,ult\ from all the valid a\sajs returned.

\ ) Cand~date B mean I 1 Laborator) j Intralaborato~ potenclei (nmolei, actrte l 1 number , 5tandard dev~ation l l ner amnoule) I I

Mean

I

Standard j CV deviation 1 (92)

1

*The results from invalid assays, indicated by brackets ( ), were not included in the calculation

Figure 6. Mean potencies of candidate B (nmoles AAT per ampoule).

AAT potency of sample B (051152)

Mean

- - - - - - - Confidence limits (95 %)

0 0 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 i 3

Lab number

WH01BSIOS.2092 Page 46

Table 7. Mean potencies of candidate G (nmoles AAT per ampoule), Potenc~e., are calculated from the mean results from all the \d id assays returned.

Figure 7. Mean potencies of candidate G (nmoles AAT per ampoule).

AAT potency of sample C (051162)

Mean

--..----- Confidence limits (95 O/O)

0 - 0 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4

Lab number

WHOBSIO8.2092 Page 47

TabIe 8. Mean potencies of candidate D (nmoles AAT per ampoule). Pcltenciei are calcutated from the mean result\ from all the valid alxayb returned.

I standard deviation deviation : (Q )

13 218.05 1 0.35 l l "13

*The results from invalid assays, indicated by brackets ( ), were not included in the calculation.

Figure 8. Mean potencies of candidate D (nmoles AAT per ampoule).

AAT potency of sample D (051172)

Mean

- - - - - - - Confidence limits (95 %)

0 0 1 2 3 4 5 6 7 S 9 1 0 1 1 4 2 1 3

Lab number

WHOIBSfO8.2092 Page 48

Figure 9. Distribution of results for AAT potency of candidate A. Columns represent the number of laborittories with results in the corresponding ,4AT concentration range. Each box represents an irrdividual laboratoq by number. u-ith the shading identifying whether the assays were kinetic or endpoint. and with a plate reader or spectrophotometer.

A (0511 50) AAT Potency - All data

Methods

100. 150. 200. 250. 300.

Potency nmoleslampoule

WHOIBSIO8.2092 Page 49

Figure 10. Distribution of results for AAT potency of candidate B. Columns represent the number of laboratories with results in the cowesponding AAT concentration range. Each box represents an individual laboratory by number, with the shading identitjling kvhether the assays were kinetic or endpoint, and with a plate reader or spectrophotometer.

B (0511 52) AAT Potency - All data

Plate - kinetic

Plate - endpoint

Spec - kinetic

0

100. 150. 200. 250. 300.

Potency nmoleslampoule

WHOIBSf08.2092 Page 50

Figure 11. Ilistribution of results for AAT potencj of candidate C. Column., repre4ent the number of labordtoriei filth reculti, in the correspond~ng )%AT concentration range Each box repremm an indil idual laboratorq by number. ith the \hading identify~ng uhether the ai,\aqs \+ere klnetlc or endpoint. and u ~ t h a plate reader or \pectrophotometer

C (0511 62) AAT Potency - All data

Methods

100. 150. 200. 250. 300.

Potency nmoleslam poule

WHOBSfO8.2092 Page 51

Figure 12. Distribution of results for AAT potency of candidate D. Columns represent the number of laboratories ~ l i t h results in the corresponding /'\AT concentration range. Each box represents an individual laboratory by number. with the shading identifying whether the assays were kinetic or endpoint, and with a plate reader or spectrophotorneter.

D (0511 72) AAT Potency - All data

Methods

0 Plate - kinetic

Plate - endpoint

Spec - kinetic

Spec - endpoint

100. 150. 200. 250. 300.

Potency nmolesiampule

MiHOIBSIO8.2092 Page 52

Tabie 9. Summary of the mean potencies for each of the candidates calculated with statistical outliers removed, Results are expressed as nmoles of active AAT per ampoule. These values are also convened to milligrams of active AAT using a molecular weight of 5 1 kDa for plasrua derived rnaterial and 43.5 kDa for recombinant, based on LC/IZfS data

i Poterlcy of actrte AXT Potency of acrr t e A.4T Csndldate / (nmole~lampoule) 1 img/ampoule)

I r 95 9 confidence ~ntert al) 1 195 5; confidence ~nter\aI)

?VHOfBS/08.2092 Page 53

Table 10. Results of the total protein concentration determined by Bradford assa using the ART reference preparation as standard. The re\uIt\ are given in term of milligrams of protein.

Table 11. Results of the total protein concentration determined by Bradford assay using laboratories' in-house standard. The results are given in terms of milligrams of protein.

Laboratory

7 &

3 3 S

3 6 9

Candidate

Standard and

Human serum 1 10.10 i

l

Albumin / 10.17 .4AT 1 1 1.33 / Albumin 9.33 1

(13.35)

C

12.70

12.30 13.36 1 1.75 10.30 13.05 ,

D

10.30

10.58 ' 11.74 641 8.5 1 13.09

Mean and C V %

(C 1

12'41 ( 10.561

,

Album~n 1 8.31 1 1 9.63 AAT !12 .19 / I 13.56

1 1.74 ' 12.85

Mean and CV 5k

(D)

10.62 (15.35) 1 1 . 1 1 ( 1 0 ? 2 )

wHOISS/OS.2092 Page 54

Table 12. Resuits of the Antigenic determination of AAT concentration using the ART referenee preparation as standard. The result\ are gi\en tn term, of miI1igrams of AAT.

Table 13. Results of the Antigenic deterdnation of AAT concentration using the Iaboratories' in-house standard. The results are giten in terms of milligmms of AAT

WHOIBSIO8.2092 Page 55

Table 14, Specific activities of the four AAT candidates by antigen. Specific activities are presented as a percentage based on the potencies of active AAT in Table 13, and the AAT contents determined by antigen using the ART reference preparation as standard (Table 16) or labctratories' usual standard (Table 17); or by told1 protein using the ART reference preparation as standard (Table 14) or laboratories' usual standard (Table 15).

WHOIBSIO8.2092 Page 56

Alpha-l Foundation kYorkshop, April 2005, Gincinnati, C'S4 Attendees of the shndardisation study steering group meeting:

WHOII3S108.2092 Page 5.7

Prof. Mark Brantly Uni\er\itj of Florida College of &Medicine Dixision of Pulmonar? & Critical Care Medicine J. Hillis iliiller Health Cenrer PO Box f 00225 Gainesvilie FL 326 10-0225

Prof. Gerald McElvaney Re\piratory Research Di\ ision. RCSl Education and Research Centre, * Beaumont Hospital, Dublin 9. Ireland

Dr Colin Longstaff Biotherapeutics, Haemostasis Section NIBSC. Blanche Lane, South Mimms. Potters Bar, HERTS EN6 3QG. UK

Dr Craig Thelwell Biotherapeutics. Haemostasis Section NIBSC, Blanche Lane. South Mimms, Potters Bar, HERTS EN6 3QG, UK

Dr Virginia B. Pate1 a Athens Research & Technology Inc PO Box 5494 183 Paradise Blvd Athens, CA 30603 Dr Afcbin Safavi Talecri5 BioTherapeutici Biological Product\. BioAnalj tics 1017 Main Campu\ Drive Suite 1200 Raleigh. KC 37606

Dr Andre% Shrake Divi5ion of HaematoIog>* mAICBER 1301 Rockl~ille Pike-HFM 333 Rock\ ills ,91D 20852- 1348

Dr Ewa h"lr\zal Division of Haen~atology FDAICB ER 1401 Rockville Pike-HFM 343 Rockville MD 20852- 1448

Symma Finn Alpha- l Foundation 2937 SW 27th Avenue Suite 302 Miami, Florida 33 133

Dr VaI Romherg ZLB Behring PO Box 511 Kankakee If, 6090 1

WHO1BS/OS.2O92 Page 58

LIST OF PARTICIPANTS

We are extremely grateful to the follouing laboratorie\ for their contribution to the r e ~ l t s of this 5tudy:

Prof. Gerald McElvaney Dr. Tornis Carroll. Respiratory Research Division, RCSI Education and Research Centre, Beaumont Hospital, Dublin 9. Ireland

Dr Petra Schulz Octapharma Pharmazeutika Produktionsges.m.b.H OberlaaerstraBe 235 1 100 Vienna Austria

Dr Silva Bergqvist Octapharma AB Nordenflychtsv. 53 1 1287 Stockholm Sweden

Dr h'lontse Costa Dr Marta Gentana Istituto Grifols. S.A. Can Gua5h. 2 08 150 Parets del Vallks Barcelona-SPAIN

Dr. Guy Rautmann Phamacop6e Europkenne Conseil de 1' Europe 223-226 Avenue de Colrnar F-67 100 Strasbourg France

Dr Chong Loh Dr Lu Liu Therapeutic Goods Administration (TGA) 136 Narrabundah Lane. Syrnonston, ACT 2609 Australia

Dr Mark Kessler Kamada, Ltd. Kibbutz Beit Kama Negev 85325 ISRAEL

Dr Virginia B. Pate1 Athens Research & Technology Inc PO Box 5494 183 Paradise Blvd Athens, GA 30603

Dr Susanne Breitner-Ruddock Paul-Ehrlich-Institute IG 7 13 Paul-Ehrlich-Str. 5 1-59 D63225 Langen Gemanq

wHO~S/OS.2092i Page 59

Aline Laulan taboratoire Fran~ais du Fractionnement et des Biotechnologies (LFB 1 BP 200659 Rue de Trevise 59 01 l LILLE GEDEX FRANCE

Dr Peter Turecek Dr Peter Matthiessen Baxrer AG

m Benatzkygasse 2-6 A- 1220 Vienna, Austria

Dr Afshin Safavi Talecris BioTherapeutics Preclinical-Clinical Assay Development. BioAnalytics 10 17 Main Campus Drive Suite 1200 Raleigh. NC27606

Dr Koen vander Drift Talecris BioTherapeutics BioAnalytics 10 17 Main Campus Drive Suite 1200 Raleigh, NC 27606

Dr Todd Willis Talecris BioTherapeutics. R&D BioAnalytics 8368 t7.S. 70 West Clayton. KC 27520

Dr David Lingenfelter Dr Sue Preston Arsiva Phamaceuticals Inc 2020 Challenger Drive Suite 101 Alarneda CA 94501 Dr Lorraine Peddada Grifolc Biologicals Inc 5555 Valley Boulevard Los Angeles CA 90032

Dr Val Romberg Dr Scott Kee ZLB Behring POBox 51 1 Kankakee IL 60901

Dr Ewa Marszal Dr Andrew Shrake Division of Haematology FDAlCBER I40 1 Rockville Pike-HFM 343 Rockville MD 20852- 1448

Dr Craig Thelwell Biotherapeurics Haernoitasis Section NIBSC, Blanche Lane. South Mimms, Potterc Bar, HERTS EX6 3QC. CK

WHOllBS108.2092 Page 60

APPENDIX C :

Overview of the calculation of % active trypsin by active-site titration. Representative active-site titration data analysed by plotting absorbance at 402 nm against time and fitting a curve describing a burst followed by steady state phase. The linear steady state phase was extrapolated back to time zero and this absorbance corresponds to the initial burst of 3- nitrophenol. In this example a burst of 0.566 (absorbance at 402 nm) is converted to a molar concentration uf trypsin by dividing the extinction coefficient 16034 (derived from the mean of all the labs). This is then expressed as a % of the total trypsin concentration. calculated from the weight of trypsin. In the study the C/c active trypsin was calculated from the results from 9 laboratories each performing 3 replicate assays with 5 different trypsin dilutions.

-.W-----

Parameter Value Std Error 6 = Apparent rate constant

= rate

lnlt~al rate 190) 0 160555 N

0 001 363 o f~nal rate iRf) 0 000190 d

0 000004

a kobs (k) 0 283084 0 002503

0 2 lnltlal absorbanceiAo)4 44299e 007 0 000691

Ro = Initial rate

0 20 40 60 80 100120140160180200 := 0.566

t !S)

Divide by molar extinction coefficient 116034)

:= 35.30 pM

IJ Tj Total Trypsin conc, iby weight) = 43.00 pM U)

S?

PVHOIBSIOS.2092 Page 61

Appendix D, Draft IFC

THE 1VHO l" ISTERNATIONAL STAXDARD FOR ALPHA-1-ANTITRYPSI?j NIBSC CODE 051162 Draft version 1.0, July 2006

'This material is not for in vitro diagnostic use"

1. INTRODUCTION

a Alpha-] -Antitrypsin (AAT) belongs to the serpin (serine proteinase inhibitor) fhmily of inhibitors and elastase is the main physiological target. Plasma-derived therapeutic AAT products are used to treat AAT deficiency (Alpha- l !. a genetic disorder identified in virtually all populations that can cause liver and lung disease in adults and children. The 1" International Standard fbr AAT is made from AAT purified from plasma, was provided by a manufacturer of therapeutic concentrates and is intended primarily for use to standardise the determination of potencies of AAT used for replacement therapy.

2. UNITAGE

The potency of the 1" International Standard for AAT was determined as past of a collaborative study where the inhibitory activity of AAT was determined by titration against trypsin. The molar concentration of active trypsin was determined by active site titration using 4-Nitrophenyl 4-guanidinobenzoate hydrochloride (NPGB). AAT and trypsin form a tight 1 : 1 stoichiometric complex so titration of AAT against a known concentration of trypsin allows expression of active AAT concentration in molar units. The amount of AAT in this International Standard is 243 nmoles per ampoule. This is equivalent to 12.4 mg active AAT per ampoule applying a molecular weight of 5 1 000 @mole. determined by mass spectrometry for this preparation.

Q International Standards are traditionally assigned units based on consensus values and represent the primary standard for a particular measurand and as such are not assigned an uncertainty. The uncertainty of the ampoule content of' 051 162 may be considered as the coefficient of variation of the fill which was 0.15%

3. CONTENTS

After reconqtitution each ampoule will contain. in addition to actite AAT, phosphate buffer and saline ( l 7 mM phosphate: 38 mhf chloride. 8 1 rnM \odium at pH 7.0) and 141 rnM mannitol added as a bulking agent. No other proteini were added. A total of 9792 ampoule\ of 051162 were prepared with a mean filling ueight of 1.0057 g (cv = 0.15 %). a mean dry weight of 0.0368 g ictr = 3.74 9;) and residual moisture of 0.6615 Cic (cv = i 1.31 55).

WHOfBSf08.2092 Page 62

4. CAUTION

THIS PREPARATION IS NOT FOR ADMINISTRATION TO HIJMANS.

The preparation containr material of human origin, which ha\ been te\ted and found negative for HBsAg. HI\; antibody, HCV antibody- and HCV RNA bq PCR.

As with all materials of biological origin, this preparation should be re~arded as potentially hazardous to health, It should be used and discarded according to your own laboratory's safety procedures. Such safety procedures probably will include the wearing of protective gloves and avoiding the generation of aerosols. Care should be exercised in opening an~poules or vials. to avoid cuts.

5. DIRECTIONS FOR OPENING THE DIN AMPOULE

DIN ampoules have an 'easy-open' coloured stress point, where the i~asrow ampoule stem joins the wider ampoule body.

Tap the ampoule gently to collect the material at the bottom (labelled) end. Ensure that the disposable ampoule safety breaker provided is pushed down on the stem of the ampoule and against the shoulder of the ampoule body. Hold the body of the ampoule in one hand and the disposable ampoule breaker covering the ampoule stem between the thumb and first finger of the other hand. Apply a bending force to open the ampoule at the coloured stress point, primarily using the hand holding the plastic collar.

Care should be taken to avoid cuts and projectile glass fragments that might enter the eyes. for example, by the use of suitable gloves and an eye shield. Take care that no material is lost from the ampoule and no glass falls into the ampoule. Within the ampoule is dry nitrogen gas at slightly less than atmospheric pressure. A new disposable ampoule breaker is provided with each DIN ampoule.

6. USE OF AMPOULED MATERIAL

This International Standard is intended to be a primary reference material for local calibrators and should not be used as a reagent. Ampoules should be stored at -20 "C or below (but are shipped at ambient temperatures and are stable for short periods). Before use. the ampoules should be warmed to room temperature and the contents of the ampoule should be reconstituted in 1 m1 of distilled water and stored on ice for use within the same day.

I Y HOISS108.2092 Page 63

7. STABILITY

It is the policy of ?VHO not to assign an expiry date to their international reference materials. They remain valid with the assigned potency and status until withdratvn or amended. Initial stability trials on material stored at elevated temperatures (4, 20. 37 and 35 "C) have shown no deterioration in actilrity relative to a reference sanlple at -20 "C over 7 months. We therefore believe the arnpouied material is stable but longer term stability trials are ongoing.

Reference materials are held at NIBSC within assured, temperature-controlled storage facilities. Reference hlaterials should be stored on receipt as indicated on the label. Once reconstituted, diluted or aliquoted, users should detemine the stability of the material according to their own method of preparation, storage and use.

KIBSC follows the policy of WHO with respect to its reference materials.

Users who have data supporting any deterioration in the characteristics of any reference preparation are encouraged to contact NIBSC.

8. CITATION

In all publication\, including data sheets, in which this material is referenced, it is important that the title of the preparation. the NIBSC code number, and the name and address of NIBSC are cited and cited correctly.

In all publications including data sheets in which this material is referenced, it is important that the WHO status of the preparation? specified by the title of the preparation. the name and address of the WHO International Laboratory for Biological Standards at NIBSC and the NIBSC code number are cited and cited correctly.

-m

9. PRODUCT LIABILITY

Information emanating from KIBSC i\ given after the exercise of all reasonable care and skill in its compilation, preparation and issue. but i \ provided uithout liability in its application and use.

This product is intended for use as a laboratory standard or reference material. It is the responsibility of the user to ensure that heishe has the necessary technical skills to detemine the appropriateness of this product for the proposed application. Results obtained from this product are likely to be dependent on conditions of use and the variability of materials beyond the control of NIBSC.

XIBSC accepts no liability whatsoever for any loss or damage arising from the use of this product, tvhether loss of profits, or indirect or consequential loss or otherwise. including. but not limited to, personal injury other than as caused by the negligence of NIBSC. In particular, NIBSC accepts no liability wharsoetrer for:-

MTHOf13S/08.2092 Page 64

( i ) reiulti obtained from this product; andfor ii i) non-delivery of goods or tor dan~agei in traniit.

In the etent of any replacement of goodi following loir or damage a cuitorner accept5 ac, a condition of recelpt oi a replacement product. acceptance of the -fact that the replacement is not to be conitrued ai an admis\ion of liability on NIBSC'i behalf.

'\;~7HOfSS/08.209;2 Page 65

10. ILIATIERIAL SAFETY SHEET

Other (cpecify) Contains material of human origin. 5ee cection 4 Handling: See caution, section 3

I Toxicological ~ro~er t i e s 1 Effects of inhalation: Not established, avoid inhalation Effects of ingestion: Not established, avoid ingestion *

I Effects of skin absomtion: Not established. avoid skin contact 1 Suggested First Aid Inhalation Seek medical advice

1 Ingestion Seek medical advice I C

Contact with eyes Wash with copious amounts of water. Seek medical advice. Contact with skin Wash thoroughlv with water. Action on Spillage and Method of Disposal Spillage of ampoule contents should be taken up with absorbent material wetted with an appropriate disinfectant. Rince area with an appropriate disinfectant followed by water.

I Absorbent materials used to treat spillage should be treated as biolttgical waste.